What problem does it solve? Maintaining a personal knowledge management system (AKNE) requires checking many health indicators—connectivity, orphan nodes, duplicate names, self-loops, source file coverage, vectors, edge formats, and more. Running scattered one-off checks produces inconsistent, non-cross-validatable evidence. This Skill consolidates the full operations lifecycle into a single comprehensive audit with a structured five-stage report. ## Core Features & Use Cases - 16-Metric Comprehensive Audit: Run scripts/akne-comprehensive-audit.py once to cover connectivity, isolated nodes, duplicate names, self-loops, source file coverage, vector counts, edge format, metadata quality, Wiki pollution, entity namespaces, and path prefixes. - Five-Stage Audit Report: Generate reports in the strict order of health diagnosis, content contradictions, research gaps, hypotheses, and fixes—contradictions must be resolved before hypotheses are generated. - Path Redundancy Resolution: Detect and normalize symbolic links pointing to the same inode (e.g., /home vs /media paths) to eliminate duplicate names and namespace confusion. - Use Case: Run a daily health check on your AKNE graph directory and source files; the Skill validates inputs first, executes the audit script once, and produces a traceable report with concrete metrics and recovery suggestions for any failures. ## Quick Start Run a comprehensive health audit on my AKNE knowledge graph and produce the five-stage audit report.
